Find the page each chapter starts on using the bookmarks or the thumbnail rail, then split at those pages. Splitting produces one file per section; if you only want a few chapters, extracting those pages is cleaner than splitting and discarding the rest.
A 400-page report where each department only needs its own section. A scanned book you want as separate chapters. A contract where the annexes should be circulated separately. The operation is simple; getting the boundaries right is the part that takes the time.
Find the boundaries first
Splitting is quick. Working out which page each chapter begins on is what people spend time on, and there are faster routes than scrolling.
Use the bookmarks if there are any
A PDF produced from a properly styled Word document usually has bookmarks generated from its headings, and each one points at a page. That is your chapter list, already made. Open the document and read the page numbers off the outline.
Use the thumbnail rail
Chapter openings look different from body pages — more white space, a large heading, often starting on a right-hand page. Scanning a grid of thumbnails finds them far faster than paging through.
If the PDF has a text layer, searching for "Chapter" or a known section title jumps straight to each occurrence with its page number. Fastest of the three when it works.
Three ways to split
At chosen page breaks
The one you want for chapters. Give it the pages where each new section begins, and you get one document per section.
Into fixed-size chunks
Every 50 pages, say. Useful for getting a huge file under an upload limit, and useless for chapters, since content does not fall into equal parts.
One file per page
Right for a stack of scanned single-page documents — invoices, receipts, certificates — that were scanned in one pass.
If you want three chapters out of twelve, splitting produces twelve files so you can throw away nine. Extracting pages gives you only what you asked for, and the original is untouched either way.
If the document is in the wrong order — a double-sided scan with all fronts then all backs — splitting will faithfully preserve the mess. Fix the order first with the page organiser, then split.
What happens to the document
Text stays text. Pages are copied, not re-rendered, so nothing becomes blurry and everything stays selectable.
Each part keeps only its own pages’ content, so the pieces are much smaller than the whole.
Bookmarks pointing outside a part become meaningless in that part — expected, and rarely a problem.
Internal links to other chapters cannot follow across files. If cross-references matter, splitting may be the wrong approach entirely.
The original is never modified.
Practical points
Check before you download
Preview each resulting document and look at its first page. An off-by-one boundary — the last page of chapter two heading chapter three — is obvious in a preview and annoying to discover later.
Take the ZIP
A split into fifteen parts means fifteen downloads. Taking them as a single archive is faster and keeps them together.
Name them as you go
Files named part-1 through part-15 are unhelpful within a week. If you are circulating them, renaming to the chapter titles is worth the two minutes — batch renaming handles it if there are many.
A note on privacy
Splitting is a structural operation a browser can do entirely on its own, which matters because the documents people split — contracts, reports, statements — are often confidential. Load the page, disconnect, and split anyway: if it works, nothing was uploaded. There is more in is it safe to use online PDF converters?.